Mistral Vibe OpenTelemetry
A practical guide to Mistral Vibe observability: OpenTelemetry tracing, OTLP export, configuration, privacy controls, and what engineering teams can learn from agent execution data.
Updated September 2026 ยท 10 min read
Mistral Vibe
CLI coding agent
OpenTelemetry
Tracing
Traces
- Agent
- Model
- Tools
Compatible backend
Your observability stack
Quick answer
What is Mistral Vibe OpenTelemetry?
Mistral Vibe supports OpenTelemetry tracing for observing AI-agent execution. When enabled, Vibe can export traces for supported agent, model and tool operations through OTLP/HTTP to compatible observability infrastructure.
These traces help engineering teams understand how an AI coding agent executes tasks and how models and tools participate in agent workflows.
In short: OpenTelemetry tracing makes supported Mistral Vibe agent workflows observable outside the coding agent itself.
Telemetry signals
What OpenTelemetry Data Does Mistral Vibe Provide?
Vibe's documented OpenTelemetry integration is built around distributed tracing of agent, model and tool operations.
- Available
Traces
Distributed traces provide visibility into supported operations across Vibe agent workflows.
- Traced
Model Operations
Observe supported interactions between Vibe and the AI models involved in agent execution.
- Traced
Tool Operations
Follow supported tool executions as they participate in agent workflows.
Mistral Vibe's documented OpenTelemetry capability focuses on tracing. Do not assume that the same native metrics and logs available in other coding agents are also exported by Vibe.
Trace reference
What Can Mistral Vibe OpenTelemetry Trace?
OpenTelemetry tracing provides a structured view of how operations relate to one another during an AI-agent workflow. Identifiers below come from Mistral's official Vibe documentation and the mistralai/mistral-vibe repository.
| Trace category | Official identifier | What it helps observe |
|---|---|---|
| Agent operations | invoke_agent | How the coding agent executes supported tasksRoot span for an agent run, carrying gen_ai.operation.name = invoke_agent, gen_ai.agent.name and, when available, gen_ai.conversation.id. |
| Model operations | chat | Interactions between the agent and AI modelsModel-call span with gen_ai.operation.name = chat, gen_ai.provider.name and gen_ai.request.model. Token attributes gen_ai.usage.input_tokens and gen_ai.usage.output_tokens are recorded on the span. |
| Tool operations | execute_tool | Tools invoked during agent executionTool span with gen_ai.operation.name = execute_tool, gen_ai.tool.name, gen_ai.tool.call.id and gen_ai.tool.type. |
| Workflow relationships | gen_ai.conversation.id | How supported operations relate within a traceVibe propagates the conversation ID as OpenTelemetry baggage so descendant spans, including spans created by the Mistral SDK, attach to the same workflow. |
| Hook operations | vibe.hook.name | Hooks that run around supported tool executionsHook spans carry vibe.hook.name and vibe.hook.type, plus the related tool attributes. |
| Execution timing | span duration | Timing information available through spansStandard OpenTelemetry span start/end timings and status codes; model-call spans also record http.response.status_code. |
Common span attributes
gen_ai.operation.nameinvoke_agent, chat or execute_tool, following the OpenTelemetry GenAI semantic conventions.
gen_ai.provider.nameProvider behind the model call, for example mistral_ai.
gen_ai.request.modelModel requested for the call; gen_ai.response.model records the responding model.
vibe.provider.api_styleVibe-specific attribute describing the provider API style used for the call.
vibe.request.streamingWhether the model call was streamed.
vibe.request.call_typeVibe-specific call type recorded from request metadata.
Use cases
What Can Engineering Teams Learn From Mistral Vibe Traces?
Traces describe how an agent run unfolds, which makes them useful for understanding execution behavior rather than counting activity.
Agent Execution
Understand how supported operations unfold as Vibe works through an AI-assisted development task.
Model Interactions
Observe where AI model operations occur within agent workflows.
Tool Usage
Understand which supported tools participate in execution and where they appear in a trace.
Performance & Troubleshooting
Use trace relationships and timing information to investigate slow or problematic operations within supported workflows.
Trace data explains execution behavior. It should not be interpreted as a standalone measure of developer productivity.
How it works
How Mistral Vibe OpenTelemetry Works
When OpenTelemetry is enabled, Mistral Vibe can generate traces for supported agent operations and export them through OTLP/HTTP to compatible observability infrastructure.
Mistral Vibe
- Agent
- Model
- Tools
OpenTelemetry tracing
Spans
OTLP/HTTP
Vibe appends /v1/traces
Collector or compatible backend
Configuration
How to Enable OpenTelemetry in Mistral Vibe
OpenTelemetry tracing must be explicitly enabled before Vibe exports trace data. Vibe is configured through config.toml, at project level or user level.
# ~/.vibe/config.toml (or ./.vibe/config.toml for a project)
# Telemetry must be enabled for OpenTelemetry tracing
enable_telemetry = true
# Export OpenTelemetry traces for agent, model and tool operations
enable_otel = true
# Base URL of your OTLP/HTTP collector. Vibe appends /v1/traces.
# Leave empty to use the configured Mistral telemetry endpoint.
otel_endpoint = "https://otlp.example.com"
# Client-side span attribute redaction: default | strict | none
otel_redaction = "strict"Keep OpenTelemetry tracing disabled, or restrict exported span content with redaction, when external observability is not required.
View advanced configuration
enable_telemetryโ Boolean, default true. Vibe requires it for OpenTelemetry tracing; disabling it also disables trace export.enable_otelโ Boolean, default false. Exports OpenTelemetry traces for agent, model and tool operations.otel_endpointโ String, default empty. Base URL of a custom OTLP/HTTP collector; Vibe appends /v1/traces. When empty, Vibe uses the configured Mistral telemetry endpoint.otel_redactionโ default redacts sensitive values, strict redacts sensitive attributes, none disables redaction.Config file locationsโ ./.vibe/config.toml for project-level settings takes precedence over ~/.vibe/config.toml.
The complete key list is documented in Mistral's Vibe configuration reference.
Privacy
Privacy and Trace Redaction
Agent traces can contain detailed contextual information about AI-assisted development workflows. Vibe provides controls that organizations should review before exporting telemetry.
Trace Content
Review which information is included in spans before sending traces to an external observability backend. Tool spans can include tool-call arguments and results.
Redaction
Use Vibe's supported redaction controls to limit sensitive information included in exported traces. otel_redaction accepts default, strict or none.
Telemetry Destination
OTLP export allows organizations to choose where supported OpenTelemetry trace data is processed and stored, using otel_endpoint.
Good practice: Apply the most restrictive redaction policy compatible with your observability requirements. Collect only the telemetry required for the questions your organization wants to answer.
From telemetry to analytics
OpenTelemetry Gives You Traces. What Do You Do With Them?
Mistral Vibe traces can reveal how agent, model and tool operations interact during execution. But raw traces are primarily technical observability data.
Understanding AI adoption at an organizational level requires connecting agent telemetry with broader patterns around usage, workflows and AI-assisted development.
Turn AI Coding Telemetry Into Engineering Intelligence
helloMetry transforms supported AI coding-agent telemetry into structured analytics designed to help engineering organizations understand how AI is being adopted and used.
- Agent Activity
- AI Workflows
- Tool Usage
- Adoption
Overview
- Code38%
- Analysis24%
- Writing18%
- Research12%
- Other8%
- /code_review142 usages4.892
- /debug98 usages3.210
- /refactor85 usages2.761
- /data_analysis74 usages2.104
- /summarize68 usages1.874
Explore Other OpenTelemetry Integrations

Claude Code
Claude Code OpenTelemetry: metrics, events and traces, configuration and privacy considerations.
Read the guideGemini CLI
Gemini CLI OpenTelemetry: available metrics, logs and traces, setup and privacy considerations.
Read the guideGitHub Copilot
GitHub Copilot OpenTelemetry: agent metrics, events and traces, setup and privacy considerations.
Read the guideCursor
Cursor OpenTelemetry: exported metrics and logs, destination setup and privacy considerations.
Read the guideQwen Code
Qwen Code OpenTelemetry: metrics, logs and traces, OTLP setup and privacy considerations.
Read the guideOpenAI Codex
Codex OpenTelemetry: metrics, events and traces, the [otel] configuration and privacy considerations.
Read the guide
FAQ
Frequently Asked Questions
Does Mistral Vibe support OpenTelemetry?
Yes. Mistral Vibe supports OpenTelemetry tracing for supported agent, model and tool operations.
What does Mistral Vibe export through OpenTelemetry?
Mistral Vibe's documented OpenTelemetry capability focuses on distributed traces representing supported agent, model and tool operations.
Does Mistral Vibe support OpenTelemetry traces?
Yes. Vibe can export supported traces through OTLP/HTTP to compatible observability infrastructure.
Does Mistral Vibe export OpenTelemetry metrics?
Mistral Vibe's documented OpenTelemetry functionality currently focuses on tracing. Refer to the latest official documentation for any changes to supported telemetry signals.
Can Mistral Vibe tool activity be observed?
Supported tool operations can appear within Vibe's OpenTelemetry traces, providing visibility into how tools participate in agent execution.
Does Mistral Vibe provide privacy controls for traces?
Vibe provides supported redaction controls that organizations can use to limit information included in exported telemetry.
Does helloMetry replace OpenTelemetry?
No. OpenTelemetry provides the standard used to generate and export observability data. helloMetry uses supported AI coding telemetry to provide engineering analytics.
Sources & further reading
Primary technical source: Mistral, for all Vibe-specific behavior
Mistral โ Vibe Code CLI configuration referenceOfficial repository
mistralai/mistral-vibeStandard
OpenTelemetry โ Traces